Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

Can autonomous vehicles make moral and ethical decisions?

Status
Not open for further replies.
Replies continue below

Recommended for you

It feels like we're coming-up on Asimov's "Three Laws of Robotics":

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.

2. A robot must obey orders given it by human beings except where such orders would conflict with the First Law.

3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.



John R. Baker, P.E. (ret)
EX-Product 'Evangelist'
Irvine, CA
Siemens PLM:
UG/NX Museum:

The secret of life is not finding someone to live with
It's finding someone you can't live without
 
The problem is with the first law... by action or inaction, a human life is lost... which one to choose, is the dilemma.

Dik
 
"If the vehicle has to decide between hitting a baby carriage or a pair of old people, it probably has been driving too fast for the conditions."

There are lots of situations that are completely outside of the car's control. It, nor you, can drive as if in any instant, there will be a runaway baby carriage, or woman being pushed in front of the car that forces that decision. An excellent example just happened: Amazing, the bus driver was able to miss hitting a woman who was intentionally pushed toward the oncoming bus, but in any number of other conditions, the woman would have likely died.

But, in general, can human drivers do this problem any better? We're talking a hypothetical case where such a decision has to be made in fractions of a second.

Machines actually have an advantage that humans don't; they can be programmed to sacrifice themselves (Law 3 above) for a human, while a human's self-preservation instinct potentially overrides other options that might avoid that dilemma, or not. There have been a couple of news items where people drowned trying to save someone else, even though they, themselves, weren't swimmers.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert! faq731-376 forum1529 Entire Forum list
 
The car self-destructs and the driver is the 'cannon fodder'... better him than the sweet little ol' lady or the baby... problem solved...

Dik
 
I did see a suggestion that given the ubiquity of mobile phones, why waste time trying to identify the targets visually when you could just 'ping' all the phones in the vicinity? Of course being squashed by an AV because your phone went flat is a bit rough, maybe this should be treated as an add-on! Watching the current computer vision toolboxes trying to work in a street market is pretty interesting, hopefully the car will be programmed to slow down when the environment is too complex.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376
 
GregLocock said:
I did see a suggestion that given the ubiquity of mobile phones, why waste time trying to identify the targets visually when you could just 'ping' all the phones in the vicinity? Of course being squashed by an AV because your phone went flat is a bit rough, maybe this should be treated as an add-on! Watching the current computer vision toolboxes trying to work in a street market is pretty interesting, hopefully the car will be programmed to slow down when the environment is too complex.

If you asked me (you didn't.) this IS. A. TERRIBLE. IDEA.

Old people and babies, who are often central to these ethical dilemma word problems, often don't carry modern phones. Whether the ethical word problem is representative of the ethical dilemmas likely to be encountered by an autonomous vehicle is another question entirely..

Either way, you don't want to make vehicle path decisions based on bluetooth connectivity.
 
As I said, it would have to be an add-on, not a replacement for vision. Success would be that cars no longer drive into pedestrians. That's about a quarter of a million deaths a year according to one biased estimate .

65% of the world's population uses mobile phones (OK that's not the same as carrying one) and 1.5 billion are sold every year. Wow.



Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376
 
Greg... you likely had no idea that the invention of the photocopy machine would lead to a huge market for filing cabinets...

Dik
 
GregLocock,

If the robot is losing control, it at least ought to be able to choose between a Ford and a Chevrolet.

--
JHG
 
I'm sure it's not so much a question of "if" as it is "how will it be legislated". In order for automakers to eventually lead to full autonomy, the algorithms will have to be based on and protected by law.

I cannot see automakers commit to 100% autonomy if they are still poised to be sued by everyone and their uncle ranging from emotional distress (I'm an animal rights activist, and the car ran over the squirrel instead of that person), to hate crimes (I think the car chose to hit Group A and not Group B because there were more people of race/ethnicity X in Group A).
 
there was an online poll(in a fully graphical interface, like a computer game) some years ago where you had to choose whom to overrun (imagine your car's brakes failed, and you can swerve in another lane but there are people crossing over both lanes).
choices were, a.o., a mother and a baby, two elder people, a robber, children, ...

your actions were "evaluated" based upon the number of kills, the ethnical background of your victims, whether or not you swerved (obeing trafic regulations, ...)
The purpose was, if I recall correctly, to measure human decision driving factors so they could be implemented in autonomous vehicles.

I slept bad after this test.

 
Status
Not open for further replies.
Back
Top